Discriminações simples simultâneas e sucessivas - na formação de classes funcionais
Canovas, Daniela de Souza
MetadataMostrar registro completo
Functional classes can be established thorough simple discrimination and repeated reversals training. However, successive reversals necessarily involve experience with the occurrence of errors, which may interfere with class formation. A previous study using this procedure indeed generated a pattern of errors that seemed to be related to the position of the visual stimuli in the experimental display and to the contingency reversals. The purpose of this study was to control for these variables in procedures designed to foster the formation of functional classes with preschool children. Experiment 1 was a systematic replication of the previous study, but used a go/no-go procedure, instead of simultaneous discrimination. In a go/no-go procedure, both the S+ and the S- stimuli are always presented in the same location, one at a time, thus eliminating the spatial location of stimuli as a potentially controlling variable. Five children first learned to discriminate between stimulus pairs (three and four pairs), and then, were exposed to repeated reversals with all stimulus pairs in each trials block. Three children showed rapid reversals, with a few errors, mostly in the initial trials of each transition; two other children developed stable error pattern. In probes for class formation, the performances of two children suggested inclusion of all stimuli in classes according to their functions (S+ or S-). The probe results for the two children who presented more errors were inconsistent with class formation. Thus, the fact that errors are unavoidably inherent to the reversal procedure remained as a possible source of interference with class formation. Experiment 2 removed contingency reversals: after learning simple and conditional interrelated discriminations, children were exposed to transfer of function tests. The S + and S- of one of three pairs of stimuli in simple discriminations were presented as samples in a matching-to-sample task with a novel stimulus pair as comparison stimuli. Six participants learned simple and conditional discriminations with fewer errors. Probe trials were inserted among trials of a baseline of simple and conditional discriminations. All children consistently responded to the novel stimulus correlated with the sample that had a previous S+ function in a simple discrimination. In conditional discriminations probes, one participant showed immediate emergence and two others showed delayed emergence of new relations. Thus, the variability in individual performances could not be attributed to errors during the learning phases. In summary, both studies showed that children aged 3 years 11 months to 4 years 11 months learned several discriminations at the same time; and some of them showed formation of functional and equivalence classes. The identification of sufficient conditions for predictable class formation by all participants remains open to further investigation, since class formation and learning set are important components of the development of relational and symbolic repertoires in children at this age.